Engineered ketocarotenoid biosynthesis in the polyextremophilic red microalga Cyanidioschyzon merolae 10D

The polyextremophilic Cyanidiophyceae are eukaryotic red microalgae with promising biotechnological properties arising from their low pH and elevated temperature requirements which can minimize culture contamination at scale. Cyanidioschyzon merolae 10D is a cell wall deficient species with a fully sequenced genome that is amenable to nuclear transgene integration by targeted homologous recombination. C. merolae maintains a minimal carotenoid profile and here, we sought to determine its capacity for ketocarotenoid accumulation mediated by heterologous expression of a green algal β-carotene ketolase (BKT) and hydroxylase (CHYB). To achieve this, a synthetic transgene expression cassette system was built to integrate and express Chlamydomonas reinhardtii (Cr) sourced enzymes by fusing native C. merolae transcription, translation and chloroplast targeting signals to codon-optimized coding sequences. Chloramphenicol resistance was used to select for the integration of synthetic linear DNAs into a neutral site within the host genome. CrBKT expression caused accumulation of canthaxanthin and adonirubin as major carotenoids while co-expression of CrBKT with CrCHYB generated astaxanthin as the major carotenoid in C. merolae. Unlike green algae and plants, ketocarotenoid accumulation in C. merolae did not reduce total carotenoid contents, but chlorophyll a reduction was observed. Light intensity affected global ratios of all pigments but not individual pigment compositions and phycocyanin contents were not markedly different between parental strain and transformants. Continuous illumination was found to encourage biomass accumulation and all strains could be cultivated in simulated summer conditions from two different extreme desert environments. Our findings present the first example of carotenoid metabolic engineering in a red eukaryotic microalga and open the possibility for use of C. merolae 10D for simultaneous production of phycocyanin and ketocarotenoid pigments.


Introduction
Microalgae are diverse photosynthetic organisms which can be found across the globe in almost every environment, having evolved the capacity for growth on carbon dioxide as a carbon source and the use of (sun)light for energy. Of the many extreme global environments colonized by algae, acidic hot-springs present one of the harshest. Nevertheless, red microalgae from the class Cyanidiophyceae thrive in water, soil and endolithic environments associated with these hot-springs at temperatures up to 56 • C and pH levels as low as 0.5 (Gross, 2000). The Cyanidiophyceae typically represent the only photosynthetic eukaryotic organisms found tolerating these extreme environments. Cyanidioscyzon merolae 10D was isolated from volcanic fields near Naples, Italy (Matsuzaki et al., 2004). It is an obligate photoautotroph with a small genome, one of the first telomere-telomere (~16 Mbp) complete genome sequences of any model species (Nozaki et al., 2007). Robust tools for genetic manipulation have been developed enabling precise homologous recombination (HR) directed by 200-500 bp targeting sequences (Fujiwara et al., 2017;Takemura et al., 2019aTakemura et al., , 2019b. As a result, Cyanidioschyzon merolae 10D has emerged as the simplest eukaryotic model cell system with a growing number of useful engineered traits (Miyagishima and Tanaka, 2021), These include the introduction of a cyanobacterial acyl-ACP reductase that resulted in increased triacylglycerol accumulation without growth inhibition (Sumiya et al., 2015) and the incorporation of a Galdieria sulphuraria sugar transporter that enabled heterotrophic growth on glucose (Fujiwara et al., 2019).
The focus of this study is the modification of native carotenoid pigment biosynthesis in C. merolae 10D. Ironically, the red microalgae are blue-green in color like cyanobacteria as they share the trait of phycocyanin use as a light-harvesting photopigment and only contain chlorophyl a. C. merolae 10D has a minimal carotenoid profile lacking alpha-carotene and lutein, it accumulates β-carotene and zeaxanthin as its terminal carotenoids and completely lacks violaxanthin and neoxanthin ( Fig. 1) (Cunningham et al., 2007). The capacity for HR transgene integration into its nuclear genome, minimal intron content, and general ease of handling make C. merolae 10D an exciting candidate for green (red) synthetic biology and metabolic engineering investigations (Lang et al., 2022;Pancha et al., 2021). Its extreme growth requirements also allow C. merolae to be cultivated with minimal risk of contamination and could be a promising host for industrial-scale algal waste-stream conversion processes (Delanka-Pedige et al., 2019;Selvaratnam et al., 2022). In addition, phycocyanin from the Cyanidiophyceae is more thermostable than that currently sourced from Arthrospira platensis (Spirulina) and is a potential valuable co-product from engineered cell biomass (Rahman et al., 2017).
Recently, advances in transgene design opened metabolic engineering in the green model microalga Chlamydomonas reinhardtii, in which native carotenoid profiles have been modified to produce the ketocarotenoids canthaxanthin and astaxanthin (Amendola et al., 2023;Lauersen, 2019;Perozeni et al., 2020). Both ketocarotenoids have value for their high antioxidant properties, application as food coloring, as well as pharmacological uses (Ambati et al., 2014). Bulk production of ketocarotenoid pigments would help drive the transition to non-toxic, natural textile dyes (Shabbir et al., 2018). Carotenoid modification in the green alga was achieved by overexpression of its native β-carotene ketolase (CrBKT) and hydroxylase (CrCHYB) in vegetative green cells where they are not naturally expressed (Amendola et al., 2023;Perozeni et al., 2020). Overexpression of CrBKT resulted in color changes of the green algal cells to brown due to global changes in pigment composition -the accumulation of orange-red ketocarotenoids and both chlorophyll a and b (Cazzaniga et al., 2022;Perozeni et al., 2020). In C. reinhardtii, CrBKT expression alone generates intermediate ketolated carotenoids from native β-carotene, zeaxanthin, and partially hydroxylated carotenoids to form canthaxanthin, intermediates, and small amounts of astaxanthin. Recently, it was shown that the hydroxylation of these to astaxanthin was enhanced by co-overexpression of CrCHYB in C. reinhardtii (Amendola et al., 2023). Here, the capacity for carotenoid engineering in the model red microalga C. merolae 10D was investigated. As part of this work, a completely synthetic plasmid toolkit was built and tested, with domestication of transcriptional elements, targeting peptides, and protein tags optimized for expression of target transgenes from either one-or twogene cassette(s) from the nuclear genome of C. merolae 10D. The green algal BKT and CHYB were optimized for the red algal nuclear genome context and expressed in fusion protein constructs from these plasmids after genomic integration in the intergenic region found in the 184-185C locus of C. merolae 10D chromosome 4. Transformants with confirmed HR integration of transgenes exhibited expression of each target product and colorimetric changes to culture pigmentation caused by ketocarotenoid accumulation which were visible by eye. The effects on cellular pigments were quantified and documented. Unlike in green algae, total carotenoids were not reduced in C. merolae 10D when ketocarotenoids were produced and these pigments did not affect cellular phycocyanin titers. Growth behaviors were investigated in optimal and modelled extreme desert environments using programmed bioreactors to show the potential for scaled cultivation concepts with engineered keto-carotenoid producing C. merolae 10D. Our results indicate that the polyextremophile is readily amenable to genetic manipulation, its carotenoid profile can be modified to generate ketocarotenoids, and future bioprocesses could harvest these separately from water-soluble phycocyanin. Here, we started with a red alga which looks cyan and used green algal carotenoid biosynthetic genes to turn it redbrown while not impacting its blue pigment composition. Our findings encourage further investigations of metabolic engineering with this promising eukaryotic photosynthetic cyan-cell chassis.
Codon optimization of coding sequences (CDSs), along with removal of unwanted restriction sites, was carried out using Geneious Prime (v. 2023.0.1; Biomatters Lt., New Zealand) and C. merolae's codon usage table found in the Kasusa database (https://www.kazusa.or.jp/codon/ cgi-bin/showcodon.cgi?species=280699). Restriction enzyme recognition sequences for the enzymes listed at the top of Fig. 2A were systematically removed from all sequences used in our one-and twocassette constructs. Thus, intermediate constructs used to create the eight constructs used here are available to speed future designs. For regulatory elements and homology arms, silent single point mutations (SPMs) were introduced manually in sequences to remove unwanted restriction sites. Modified promoter and terminator sequences were analyzed and compared to original sequences using Softberry Nsite(M)-PL (www.softberry.com) and Geneious DNA-fold (v. 2023.0.1; Biomatters Lt., New Zealand) to ensure conserved regulatory motifs and secondary structures, respectively, were not altered. All SPMs were documented and are indicated in sequences as lower-case bases (Supplemental Data S1). In silico assembly and de novo synthesis of transformation plasmids using pBluesript II KS (+) (Stratagene, USA) as the backbone vector was done in the Snapgene (software v. 6.4; www.sna pgene.com) and using GenScript services (GenScript Inc., USA), respectively ( Fig. 2 and Supplemental Data S4). The full, annotated sequences of all plasmids used in this work are provided in Supplemental Data S4, which, when opened with a plasmid map program, will generate annotated plasmids maps for all plasmids. All plasmids were transformed into chemically competent E. coli JM109 cells and plasmids were extracted using ZymoPURE II midiprep kits (Zymo Research group, California).

C. merolae 10D transformation
To prepare linear DNA fragments for transformation, PCR was performed using primer set 1 (detailed in Supplemental Fig. S1 and Data S2) and plasmid DNA. The resulting products were then purified by ethanol precipitation. PEG-mediated transformation of C. merolae 10D was carried out using 4 μg of linear DNA as previously described (Fujiwara et al., 2013(Fujiwara et al., , 2021 with some modifications. Transfected cells were transferred into 8.0 mL of MA2 media in 20 mL culture tubes and allowed to recover while rotating (~80 rpm) in the outer rim of a tissue culture roller drum (New Brunswick; model TC-7; Eppendorf, USA) housed in an Algatron® incubator (Photon Systems Instruments, Czech Republic) supplemented with 3% CO 2 in air with continuous illumination (100 μmol m − 2 s − 1 ) at 40 • C for two days. Cells were subsequently collected by centrifugation, supernatant discarded, and cells resuspended in MA2 (~400 μL).
Cell suspensions were then serial diluted in MA2 and 200 μL of the dilutions were amended to equal volume of 40% corn starch with chloramphenicol ("Cm" 300 μg/mL  Fig. S1 and Data S2) to test for integration of our cassettes into the targeted neutral site. Positive transformants were then scaled up as shown in Fig. 2C and characterized via PCR, flow cytometry, fluorescent microscopy, UV-vis spectrophotometry, thin layer chromatography (TLC), and high-performance liquid chromatography (HPLC). A subculture from each was cryopreserved in DMSO for long term storage (as described above).

DNA extractions and PCR assays
Cultures were harvested by centrifugation (5 min at 14,000×g) and total genomic DNA was extracted from algal cell pellets (~50-100 mg) with a Zymo Quick-DNA fungal/bacterial extraction kit (Zymo Research group, USA) according to the manufacturer's protocol. DNA extracts were quantified using a NanoDrop One spectrophotometer (Thermo Fisher Scientific, USA). The high fidelity PrimeSTAR GXL DNA Polymerase (Takara Bio Inc., Japan) and the Hot start GoTaq polymerase (Promega Corporation, USA) were used for PCR according to the manufacturer's protocols. The former was specifically used with primer set 1 to amplify the insert DNA (HR-L to HR-R) for transfection and to screen transformants for presence of the insert at the target neutral site. All primers used to screen cultures were synthesized by IDT (Integrated DNA Technologies Inc., San Diego) and primer sequences along with PCR conditions and relative primer annealing sites are shown in Supplemental Data S2 and Fig. S1, respectively.

UV-Vis spectrophotometry
A HACH DR5000 UV-Vis spectrophotometer was used to monitor culture growth by measuring the optical density at 750 nm and to analyze pigment extracts, unless otherwise stated. In vitro spectral profiles of wild type and transformed cells were obtained using a SpectraMax i3 plate reader (Molecular Devices, CA, USA) across a range of wavelengths spanning from 300 to 850 nm.

Epifluorescence microscopy
Cells were visualized and imaged with 100X objective lens and immersion oil using an Olympus B×51 fluorescence microscope equipped with a Canon EOS RP DSLR camera. Fluorescence microscopy was performed on transformants specifically expressing the mVenus (YFP) transgene to verify localization of YFP in the chloroplast and evaluate cassettes with YFP fusions. Two different excitation filters were used for detecting pigment and YFP fluorescence: U-MWG2 and FITC-3540 B-OMF, respectively.

Flow cytometry
Flow cytometric analyses of wild type and YFP transformant cells were performed using a Guava® easyCyte™ HT BGV flow cytometer (Luminex Corporation, Austin, TX, USA) equipped with a blue (488 nm) laser; which was used to measure size (forward scatter), granularity (side scatter), chlorophyll fluorescence (692/40 nm) and YFP fluorescence (575/25 nm). All samples were normalized to 0.01 OD 750

Fig. 2. Plasmid design, culturing systems and transgene integration.
A -Synthetic plasmids were designed in silico and constructed de novo for integration of transgenes into the 184C-185C locus (HR-L and -R) on C. merolae chromosome 4. Two template plasmids were synthesized: a two-cassette (upper) and a single cassette (lower), both with chloramphenicol (CAT) resistance marker as a selection/fusion partner. Expression elements and gene fragments are separated by nonredundant restriction endonuclease sites as illustrated. pCPCCphycocyanin-associated rod linker protein promoter, CTP CMH166C -DNA Gyrase B chloroplast targeting peptide, mVenusyellow fluorescent protein reporter, StrepII -C-terminal peptide tag with stop codon, tNOSnopaline synthase terminator, pAPCCallophycocyanin-associated rod linker protein promoter, CTP CMO250C -allophycocyanin-associated rod linker protein chloroplast targeting peptide, FLAGpeptide tag with stop codon, tβ-tub -C. merolae β-tubulin terminator CMN263C. B -C. reinhardtii β-carotene ketolase (CrBKT) and β-carotene hydroxylase (CrCHYB) transgenes were codon optimized for C. merolae nuclear genome expression based on amino acid sequences and native targeting peptide removal and subcloned into either of the above two plasmids as illustrated for expression as either target-mVenus or -CAT fusion proteins. Ctransformation of C. merolae, recovery of colonies in starch spots on chloramphenicol selection, and seed train for experiments. Dpolymerase chain reaction confirmation of plasmid integration at the 184-185C neutral locus, presence of transgenes, and unialgal status (RuBisCO HaeIII digestion). Information on primers and PCR assays found in Supplemental Fig. S1 and Data S2.
(~350-450 cells μL − 1 ) and a total of 10,000 events were recorded per sample. Data acquisition and analysis was done using GuavaSoft v. 3.4 software (InCyte; Luminex Corporation).
For the Algem photobioreactor growth experiment, the cell densities were measured using an Invitrogen Attune NxT flow cytometer (Thermo Fisher Scientific, UK) equipped with a Cytkick microtiter plate autosampler unit as recently described . Each sample was diluted 1:100 with 0.9% NaCl solution and loaded into a 96-well microtiter plate in technical triplicates, the cell density was measured from this plate using the autosampler. Samples were mixed three times immediately before analysis, and the first 25 μL of the sample was discarded to ensure a stable cell flow rate during measurement. For the data acquisition, 50 μL from each well was analyzed.

Biomass determination
For the 20 mL culture tube growth experiment, ash-free dry weights (AFDW) were determined using OD 750 values and an OD 750 to AFDW correlation coefficient, which was determined for each transformant prior to the experiment and found to be the same for all strains: AFDW (g/L) = 0.27 * (OD 750 nm). This correlation coefficient was determined as previously described (Dandamudi et al., 2021). For Algem photobioreactor growth experiments, biomass was measured by vacuum filtration of 4 mL of each test on pre-weighted filters (0.45 μm). The algal cells were dried at 60 • C for 24 h in petri dishes, then allowed to cool before weighing the filter with the biomass. All measurements consisted of technical and biological triplicates.

Pigment extraction and analysis
All extractions and analyses of pigments were carried out in dark or dim light to avoid photodegradation. For phycocyanin extraction, 4.5 mg of freeze-dried biomass was added into 1.5 mL 0.1 M phosphate buffer (pH 7.0) and subjected to bead beating (Bullet Blender® STORM 24, Next Advance, USA) using a mix of 0.15 mm and 0.5 mm zirconium oxide beads at the highest speed for 5 min. The supernatant was recovered by centrifugation at 12,000×g for 5 min, and the pellet was reextracted under the same conditions. Both supernatants were combined and analyzed spectrophotometrically.
The extraction of carotenoids and chlorophyll a was performed using 10 mg of freeze-dried biomass added to 800 μL of acetone containing 0.1% (w/v) butylated hydroxytoluene to prevent carotenoid oxidation. The mixture was homogenized via bead beating as described above. The supernatant was collected after centrifugation at 12,000×g for 3 min, and the remaining pellet was subjected to three additional extractions using 600 μL of acetone until the supernatant became colorless. All the supernatants were pooled and evaporated to dryness under a stream of nitrogen. For carotenoid saponification, dried extracts were resuspended in 300 μL ethyl acetate and treated with 300 μL 5% (w/v) methanolic KOH under constant shaking at room temperature for 2 h. To stop the reaction, 100 μL of 10% (w/v) NaCl, and 200 μL of deionized water were added to the reaction mixture, and carotenoids were extracted four times with hexane:MTBE (1:1, v/v, 300 μL per extraction) using centrifugation (12,000×g, 1 min) to separate the layers. The organic layers were collected and combined, then evaporated to dryness under a stream of nitrogen. Dried extracts, whether saponified or nonsaponified, were dissolved in 1 mL of acetone, filtered using a 0.45 μm nylon filter in preparation for pigment analysis by TLC, UV-Vis spectrophotometry and HPLC.
TLC was used to separate and identify carotenoids. 20 μl aliquots of the pigment extracts and carotenoid standards were spotted on 20 × 20 cm pre-coated silica gel 60 TLC plates (Supelco) and eluted with a mobile phase of hexane:acetone (7:3, v/v). The concentrations of phycocyanin, chlorophyll a and total carotenoids were determined spectrophotometrically. The absorbance of phycocyanin extracts was measured at 620 and 652 nm, and the concentration of phycocyanin was calculated using previously published equations (Bennett and Bogorad, 1973). For the assessment of chlorophyll a and total carotenoid contents, absorbance of extracts was recorded at 662 and 470 nm, respectively, and the concentrations of chlorophyll a and total carotenoids were calculated according to previously published equations. Separation of carotenoids and their quantification were conducted by reverse-phase HPLC (Waters Alliance 2695 Separations Module coupled with a 2996 photodiode array detector) as described in (Amendola et al., 2023;Perozeni et al., 2020). The HPLC system was equipped with a C18 column (Waters Spherisorb ODS2 Column 5 μm, 4.6 mm × 250 mm, Supelco, Inc., Belefonte, PA, USA) and a 15 min gradient of ethyl acetate (0%-100%) in acetonitrile-water-triethylamine (9:1:0.01, v/v/v) was employed at a flow rate of 1 mL/min. Carotenoid peaks were identified by comparing retention times and spectra to carotenoid standards, which were also used to quantify carotenoids using standard curves (Supplemental Data S3). of 3 sets were prepared: one for daily growth metrics and the other two for pigment analysis. The culture tubes were arranged in the outer rim of a tissue culture roller drum that was housed in an Algatron® incubator as above, according to their respective light conditions. Water acidified to medium pH was added as need to account for evaporative losses.
When sampling daily for growth metrics (≤100 μL), the same volume that was removed for sampling was replaced with medium. Culture density was monitored spectrophotometrically (as described above). The two sacrificial sets of tubes for pigment analysis were collected at different growth phases: one at log phase and the other at stationary phase. Biomass was collected from each culture tube by centrifugation (4,200×g for 10 min) and pellets freeze dried for pigment analysis. Growth metrics and pigments analysis (N = 3x biological and technical replicates) was done as described in the previous sections.

Algem photobioreactor performance benchmarking in modelled environments
C. merolae 10D WT and transformant lines ii and viii were first precultured in MA2 liquid medium (pH 2.3) in 125 mL Erlenmeyer flasks with a working volume of 10 mL for 4 d under continuous agitation (100 rpm) and illumination (90 μmol photons m − 2 s − 1 ) in a CO 2 (4%) incubator at 42 • C. These cultures were then used to inoculate 1 L Algem photobioreactor flasks (working volume 400 mL) with a target density of 3 × 10 6 cells mL − 1 . To simulate outdoor light and temperature conditions of Thuwal, Saudi Arabia (22.3046 N,39.1022 E) and Mesa, Arizona, United States (33.305130 N, − 111.67300 W), environmental conditions were developed based on data sets reported by  and obtained from the AzCATI facility, respectively. Four different conditions were used to evaluate the growth performance of each strain: (1) constant light (1500 μmol m − 2 s − 1 ) and temperature (42 • C), (2) 12:12 h light:dark with these same light and temperature conditions, and simulated seasonal environmental conditions of (3) Thuwal and (4) Mesa, with the months of February, May, August, and November representing winter, spring, summer, and autumn (respectively). Samples of 15 mL were collected daily for cell concentration, biomass quantification, and carotenoid analysis as described above. The same volume that was removed for sampling was replaced with sterilized water acidified to medium pH.

Statistical analysis
Data analysis was conducted using both biological and technical triplicates. The technical replicates were used to calculate the mean values for each of the biological replicates and were then used to calculate standard errors using Microsoft® excel software. Error bars, representing the standard error, are displayed in relevant graphs.

Results and discussion
The polyextremophile C. merolae 10D is restricted to low pH (0.5-5) and temperatures from 35 to 56 • C (Miyagishima and Tanaka, 2021). It has a simplified natural carotenoid profile which lacks the alpha-branch of carotenoid biosynthesis and has only β-carotene and zeaxanthin as terminal carotenoids (Fig. 1 (Cunningham et al., 2007),). In higher plants and green algae, alpha-carotene is converted into lutein, and zeaxanthin is used to create violaxanthin and neoxanthin as part of the photoprotective/photoresponsive xanthophyll cycle (Goss and Jakob, 2010;Latowski et al., 2004). As these pigments are absent in C. merolae, it is an interesting species with a simplified carotenoid substrate and biosynthesis enzymatic landscape in which to attempt carotenoid metabolic engineering. C. merolae also uses phycocyanin as a light harvesting pigment (Lang et al., 2022;Parys et al., 2021), a different photosystem structure than in green algae and higher plants, opening the question what effects carotenoid modifications would have in this system.
Ketocarotenoids are orange-red pigments that are formed through the ketolation of the terminal rings of β-carotene and zeaxanthin to form a range of intermediates towards canthaxanthin (dual-ketolated β-carotene) and astaxanthin (dual-ketolated and hydroxylated β-carotene) (Fig. 1) (Perozeni et al., 2020). Canthaxanthin and astaxanthin are formed in a range of organisms including algae, plants, bacteria and fungi (Seybold and Goodwin, 1959;Wan et al., 2021;Zhang et al., 2020). These pigments have various applications from food colorants, aquaculture feed enhancements, medicinal treatment of skin diseases, as specialty chemical conjugants, and are considered powerful antioxidants (Ambati et al., 2014). Recent reports have shown that it is possible to leverage gene redesign and synthetic overexpression of the native β-carotene ketolase (CrBKT) and hydroxylase (CrCHYB) of the green microalga C. reinhardtii to produce canthaxanthin, intermediate ketocarotenoids, and astaxanthin in this photosynthetic microbe (Amendola et al., 2023;Cazzaniga et al., 2022;Perozeni et al., 2020). The BKT adds ketone groups to the terminal rings of both zeaxanthin and β-carotene, while CHYB adds hydroxyl-groups to β-carotene ( Fig. 1) (Amendola et al., 2023). As both β-carotene and zeaxanthin are the terminal carotenoids within C. merolae and its growth conditions minimize risk of contaminating organisms, we reasoned it could be an efficient cell chassis for metabolic engineering and biotechnological ketocarotenoid production.

Synthetic transgene expression cassette design and transformation
Recent reports indicated the possibility of nuclear transformation and efficient transgene integration by homologous recombination (HR) in C. merolae (Fujiwara et al., 2013(Fujiwara et al., , 2017(Fujiwara et al., , 2019(Fujiwara et al., , 2021Minoda et al., 2004;Takemura et al., 2018Takemura et al., , 2019a. Here, it was investigated whether a synthetic-biology strategy could be used to enable heterologous expression of the green algal ketocarotenoid-biosynthesis enzymes in C. merolae. Promoter, terminator and plastid targeting signals (Miyagishima and Tanaka, 2021) were used to drive expression of C. merolae codon optimized sequences coding for CrBKT and CrCHYB in silico ( Fig. 2A) and the expression cassettes commercially synthesized de novo. The expression cassettes were designed to be modular, with each element separated by unique restriction endonuclease sites and a previously demonstrated target for HR was chosen, the 184-185C locus found on C. merolae 10D chromosome 4 (Fujiwara et al., 2017). Coding sequences for each target transgene were optimized for the C. merolae codon usage bias before synthesis and selection was achieved with a codon optimized chloramphenicol resistance (CAT) marker. Plasmids were built to express CrBKT and CrCHYB in various fusion constructs to either the CAT resistance marker or yellow fluorescent protein (mVenus, YFP) in different combinations of gene cassettes (Fig. 2B). Full, annotated sequences of all plasmids are provided in Supplemental Data S4.
To enable expression of the CrBKT and CrCHYB, different genetic fusion constructs were used to allow selection for expression with either antibiotic resistance or visually through fluorescence screening (Fig. 2B). Plasmid i was designed to express the chloramphenicol resistance marker (CAT) and localize it to the algal plastid with a targeting peptide of a native protein. Transformants generated with this act as controls for other constructs. Similarly, construct iv serves as a control for the fluorescent reporter mVenus (YFP), which was also targeted to the algal plastid through a separate targeting peptide than the CAT resistance marker (Supplemental Fig. S2). CrBKT has been shown to be a highly active enzyme in the production of ketocarotenoids and is effective in direct fusion to the spectinomycin resistance marker in C. reinhardtii (Amendola et al., 2023;Cazzaniga et al., 2022). We emulated this strategy of selection marker fusion to the CrBKT here (constructs ii, ii, vi, vii, viii) with CAT as this selection marker functions to yield resistance colonies in C. merolae 10D and also functions when localized in the algal plastid where carotenoid biosynthesis occurs . Fusion to a reporter protein can also increase the half-life of target recombinant proteins in cells and improve overall to target product yields in metabolic engineering efforts (Cheah et al., 2022). This strategy has been effective in overcoming nuclear transgene expression limitations in green algae, and was shown to be the most effective strategy for CrBKT fusion in its original report (Lauersen, 2019;Perozeni et al., 2020). Therefore, construct v was designed to express CrBKT in fusion with YFP to determine if it was more effective than CAT fusion. CrCHYB was shown to express well in C. reinhardtii where it catalyzes hydroxylation of ketocarotenoids to astaxanthin (Amendola et al., 2023). Here, we chose to attempt its expression alone (vii), in fusion with YFP (viii), or in longer fusion to the C-terminus of CrBKT-CAT (iii). Each was investigated to determine whether binary cassettes of larger sizes could be integrated into the genome by HR, and whether different efficacy in astaxanthin biosynthesis could be achieved with different fusion orientations (Fig. 2B).
Synthetically designed plasmids were used as templates for PCR to generate linear DNA fragments used in PEG-mediated transformation of C. merolae. Colonies of C. merolae 10D resistant to chloramphenicol could be readily achieved in starch beds following reported protocols  for every construct designed in this work (Fig. 2C).
Colonies were isolated by picking and grown in 400 μL MA2 liquid medium in standing glass vials prior to further analysis. For each plasmid construct, several dozen colonies were selected and checked for integration by PCR using primers listed in Supplemental Data S2. Representative clones were used to show profiles of PCR products indicating genomic integration markers (Fig. 2D) and representatives from each transformant pool used in carotenoid analysis. Expression success is described in the following section in relation to effects on carotenoid biosynthesis.

β-carotene ketolase and hydroxylase generate ketocarotenoids in C. merolae 10D
All carotenoid modifying enzymes were successfully expressed from our synthetic transgene constructs in C. merolae 10D and caused changes to the native carotenoid profiles in each strain (Fig. 3). This effect was visible already in cultures to the naked eye (Fig. 3A) and was confirmed by spectrophotometric scans (Fig. 3B) similar to those previously reported for CrBKT and CrCHYB expression in Chlamydomonas (Amendola et al., 2023;Perozeni et al., 2020). It was observed here that all C. merolae 10D transformants with CrBKT or CrBKT + CrCHYB expression exhibited a visible color change relative to the parental strain (Fig. 3A). Absorbance measurements revealed a shoulder at ~500 nm, a phenotype previously reported in organisms accumulating ketocarotenoids (Fig. 3B).
TLC of acetone extracts then indicated the presence of orange-red pigments in transformants of each construct, which were absent from the parental or control strains expressing the CAT resistance alone or CAT and YFP alone (Fig. 3C, plasmids i and iv). Transformants expressing variations of CrBKT (plasmids ii, v, vi) were observed to accumulate canthaxanthin and adonirubin as major ketocarotenoids, with minor bands of astaxanthin (Fig. 3C). The native carotenoid pathway contains hydroxylation activity to convert β-carotene into zeaxanthin (Fig. 1). However, the accumulation of mostly canthaxanthin and adonirubin in CrBKT expressing transformants indicates that the native CrtR activity does not outcompete the CrBKT activity on β-carotene substrate and is not so highly active as to further hydroxylate these ketolated products. This is similar to the native CHYB activity in Chlamydomonas, which only creates significant titers of astaxanthin when overexpressed in the green alga as well (Amendola et al., 2023). Those transformants with co-expression of CrCHYB with different fusion partners as well as CrBKT (ii, vii, viii) exhibited minor bands of these two ketocarotenoids and astaxanthin as the major band in TLC (Fig. 3C). Transformants of plasmid iii where CrBKT and CrCHYB are in a single fusion with each other, exhibited an intermediate phenotype, where astaxanthin was the major product, however, not as strong as with the two separate cassette expression in plasmids vii or viii. Patterns could be observed in non-saponified and saponified samples (Fig. 3C, upper and lower panels, respectively). These patterns were true across individual transformants analyzed in a larger TLC with two representative transformants per plasmid construct as shown in Supplemental Fig. S3. We chose only constructs ii and viii for HPLC quantification studies as these were representative of one or two transgene expression conditions and exhibited clear colorimetric phenotype, spectrographic, and TLC pattern changes (Fig. 3).
Transformants were also subjected to flow cytometry analysis to determine the expression level of fusion reporter proteins, which confirmed the strength of expression for some constructs (Supplemental Fig. S2). Plasmids iv and vi were shown to have fluorescence patterns distinct to those transformed with constructs harboring YFP fusions and those not harboring the YFP reporter (Supplemental Fig. S2). Microscopy also confirmed localization in the chloroplast via YFP fluorescence (Supplemental Fig. S2).
To determine the exact amounts of each carotenoid in the biomass, the parental strain and one transformant from plasmid ii (CrBKT) and viii (CrBKT + CrCHYB) were subjected to pigment quantification by HPLC at 437 and 465 nm (Fig. 3D-F; Table 1). Drastic differences in carotenoid profiles can be observed in the CrBKT and CrBKT + CrCHYB expressing transformants. The CrBKT expressing transformant exhibited a 33-61% reduction in β-carotene content accompanied by the disappearance of peaks #4 and 7 (zeaxanthin and β− cryptoxanthin, respectively) with the emergence of two predominant peaks #2 and 3 corresponding to adonirubin and canthaxanthin, respectively, and small amounts of astaxanthin (peak #1). The CrBKT + CrCHYB expressing transformant consequently exhibited reductions in peaks #2 and 3 and significant increase in astaxanthin content (Fig. 3D-F; Peak #1).

Fig. 3. C. merolae culture phenotypic changes and carotenoid profiles of transformants expressing different combinations of CrBKT and CrCHYB.
A -Cuvettes containing 1 mL of C. merolae transformant culture for one representative of each confirmed plasmid transformation. Babsorbance spectra of cultures pictured above, shoulder of ketocarotenoid absorbance indicated with a black arrow. C -Acetone extract TLC of one confirmed representative C. merolae transformant for each indicated plasmid with carotenoid standards. Aboveraw acetone extracts, belowsaponified extracts. Arrows indicate 1astaxanthin, 2 -adinorubin, 3 -canthaxanthin, 4 -zeaxanthin, 5 -chlorophyll a, 6 -echinenone, 7 -β-cryptoxanthin, 8 -pheophytin a, 9 -β-carotene. HPLC profiles of carotenoids from parental C. merolae 10D (D), and transformants expressing CrBKT-ii (E) or CrBKT + CrCHYBviii (F).  The cellular localization(s) of the ketocarotenoids produced in our study are not yet known. Some localization in photosystems I and II complexes is expected as native carotenoid pathways are within the plastid. Localization in lipid bodies is possible, however, unlikely as C. merolae 10D is known to make small amounts of lipid bodies, ~0.04 wt%, in nutrient replete conditions as used here (Fukuda et al., 2018). Given that our ketocarotenoid producing strains accumulated 0.9 to 1.1 wt% total carotenoids, it is difficult to imagine TAG bodies being present at levels sufficient to constitute a significant physical sink. The mechanism of ketocarotenoid localization to lipid bodies has only been described in green alga systems with specific evolved adaptation for this as a light protection mechanism (Ota et al., 2018). It is, therefore, unlikely this trait is occurring in the engineered red alga. The keto-carotenoids could also be bound to carotenoid binding proteins as described in other algal systems (Slonimskiy et al., 2022;Toyoshima et al., 2020). The precise localization of ketocarotenoids and interactions with native cellular structures will be the subject of follow-up investigations but are beyond the scope of this report.

Presence of ketocarotenoids improved total carotenoid content but slightly reduces growth rates of C. merolae 10D
Previous reports of ketocarotenoid biosynthesis in a green microalga indicated a global reduction of carotenoids and chlorophylls in transformants expressing CrBKT but increased resistance to reactive oxygen species and high-light conditions (Amendola et al., 2023;Cazzaniga et al., 2022). It was unclear how ketocarotenoid presence would affect the photosystems of C. merolae here because these photosystems also contain phycocyanin as a light harvesting pigment and exhibit a natively minimal carotenoid profile lacking alpha carotenoids and terminal xanthophylls (Cunningham et al., 2007). The transformants and parental strain were subjected to a 12-day growth experiment in 20 mL culture tubes (1.6 cm diameter) with 8 mL working volume that enable high-light penetration into the culture. Cultures were subjected to either 750 or 1172 μmol m − 2 s − 1 light intensity in a CO 2 rich (3%) environment and sampled daily for growth metrics in addition to pigment quantification at the beginning, mid (Llog), and end of cultivation (Sstationary) (Fig. 4A-H; Supplemental Data S5).
In these optimized conditions, where light penetration into the thin culture tubes and CO 2 are not limited, all cultures accumulated high rates of biomass over the 12-day period. C. merolae 10D achieved ~12 g L − 1 and ~16 g L − 1 in 750 and 1172 μmol m − 2 s − 1 , respectively (Fig. 4B).
Growth behavior of both CrBKT and CrBKT + CrCHYB transformants were ~10 and 13-14 g L − 1 , respectively, in the two light conditions ( Fig. 4A and B). Phycocyanin content per cell was not significantly different between transformants and the parental strain in either illumination condition (Fig. 4C). Total phycocyanin content was reduced in higher light conditions across all strains. Chlorophyll was overall lower in the higher light condition (Fig. 4D), while total carotenoids were lower in the parental strain, but not in transformants (Fig. 4E). Both types of pigments showed variation among the cell lines, with both ketocarotenoid accumulating strains exhibiting approximately 0.6-1.9 wt % chlorophyll a and approximately 0.9-1.1 wt % total carotenoid content ( Fig. 4 D and E).
The carotenoid profiles of each strain were unique, as shown in Figs. 3 and 4F-H, and trends observed in carotenoid species during the log phase were largely maintained in stationary phase for all cultures (Fig. 4F-H). For the wild-type 10D, zeaxanthin was the most abundant carotenoid (0.35-0.58 wt %) with β-carotene as the second most abundant (0.09-0.24 wt %, Fig. 4F). In the CrBKT expressing strain, canthaxanthin was the most abundant carotenoid (0.39-0.60 wt %, with adonirubin (keto group on both terminal rings and single ring with hydroxyl group) the second most abundant (0.29-0.37 wt %, Fig. 4G). In the CrBKT + CrCHYB expressing transformant, astaxanthin was the major carotenoid species at 0.45-0.69 wt %, with adonirubin, canthaxanthin, echinenone, and β-carotene present but much less abundant ( Fig. 4H).
The results suggest that the total carotenoid per biomass in variable light conditions seems to be relatively constant despite overall reductions in other photosystem pigments in the ketocarotenoid producing transformants. Higher-light intensities reduced overall cellular phycocyanin contents, as expected based on previous reports of the behavior of this pigment in other organisms, where it is accumulated to assist photon capture in lower-light conditions (Chen et al., 2010). Similarly total chlorophyll reduction is also observed in higher light intensities, however, the red alga is unusual to what is observed in plants and green algae in that it does not have reduced overall carotenoids when ketocarotenoids are produced and in higher light (Cazzaniga et al., 2022;Perozeni et al., 2020). This could suggest that C. merolae is a promising chassis for tailored carotenoid production, especially considering it lacks a cell wall which enables simple carotenoid extraction. Concepts which aim to concomitantly acquire phycocyanin pigment and carotenoids from the same culture could use higher-light intensities to accumulate biomass as shown here and a period of lower light intensity before harvest to increase cellular phycocyanin yields, however, such tests were beyond the scope of this work.

Modeling C. merolae growth in extreme environments
As a polyextremophile, C. merolae 10D can be grown in temperatures above cultivation norms for other algal species and in a very low pH to largely prevent contamination (Miyagishima and Tanaka, 2021). As the growth test performed in Fig. 4 was performed in small culture tubes to ensure high-light penetration, we were curious how the parental and transformant strains would perform in larger culture volumes, where light penetration would become limiting. We grew the wild-type C. merolae 10D, CrBKT (ii), and CrBKT + CrCHYB (viii) strain in 400 mL cultures using a suite of photobioreactors to tightly control environmental parameters while tracking growth (Fig. 5). Cell lines were grown at 42 • C with constant 1500 μmol photons m − 2 s − 1 illumination or 12:12 day:night light cycling to represent controlled bioreactor cultivations in optimal conditions. In addition, we used collected weather data generated on the mid-Red Sea coast (Supplemental Data S6) and in Mesa Arizona (Supplemental Data S7) to generate 8-day cultivation programs representing one month of each season in these locales. The summer months in both geographies exhibit high temperatures, with Mesa having higher midday temperatures and greater fluctuations between day and night (Fig. 5A-D, right panels).
In all bioreactor conditions, the ketocarotenoid producing transformants exhibited slightly lower optical and cell densities, as well as biomass compared to their parental strain (Fig. 5A-D). Both transformants performed similarly, suggesting that the presence of ketocarotenoids at all, rather than a specific type, caused this growth behavior difference. In continuous illumination, the 400 mL cultures achieved ~2.5 g L − 1 biomass in 6 d, while the ketocarotenoid transformants accumulated ~2.2 g L − 1 (Fig. 5A-D). Overall cell densities exhibited similar amounts in both geographies, with mid-Red Sea coast having slightly higher biomass accumulated than in Mesa (Fig. 5A-D). The higher temperatures observed in Arizona summer exceeded the capacity of the bioreactor (+50 • C), temperatures which would likely be detrimental to many algal species in culture (Fig. 5). Nevertheless, it was still possible to grow both transformed and parental C. merolae in this condition where they accumulated the ketocarotenoid products (Fig. 5E, pictures). All data for phycocyanin, chlorophyll, and carotenoid accumulation can be found in Supplemental Fig. S4.

The value of C. merolae 10D as a host for engineered carotenoid biosynthesis
The Cyanidiophyceae are polyextremophilic red algae which have emerged in recent years as interesting alternatives to other algal systems (Lang et al., 2022). Within this class are several species that are found in acidic hot springs and thrive between pH 0.5-5 and temperatures from 35 to 56 • C. These growth conditions set the Cyanidiophyceae apart from other algae in that few organisms can grow in such conditions and contamination at scale can be largely prevented. Galdieria sulphuraria is another species within this class that has been shown to be capable of rapidly becoming the dominant organism when grown directly in acidified municipal effluent (Henkanatte-Gedera et al., 2015. C. merolae 10D is an obligate phototroph and can only consume CO 2 as a carbon source (Miyagishima and Tanaka, 2021). It is also tolerant to very high levels of CO 2 gas, ammonium concentration in its medium, and high temperatures Miyagishima and Tanaka, 2021). These features potentially mean C. merolae 10D could be coupled to post-treatment high-strength wastewater polishing and industrial CO 2 emissions sources in extreme conditions such as those in desert environments modelled here.
C. merolae 10D is also interesting for biotechnological applications owing to its lack of cell wall and range of native natural products which can be rapidly separated in various phases of extraction. The cell itself contains a small lipid fraction, starch, and β-glucan in addition to ~50% protein content (Miyagishima and Tanaka, 2021). C. merolae accumulates the photopigment phycocyanin which is extractable in the water-soluble fractions of biomass and more thermal stable than that currently used in industry produced by Arthospira platensis (Rahman et al., 2017). The parental strain also accumulates large fractions of zeaxanthin and β-carotene which are both valuable hydrophobic pigments (Figs. 3 and 4). The cell, therefore, is a natural candidate for biorefinery concepts, as PC and water-soluble proteins can be readily extracted from cell-wall-less biomass and carotenoid pigments isolated from the residual water-insoluble fraction. Separate fractionation of starch and β-glucans may also be possible with appropriate bioprocess designs. This concept is illustrated in Fig. 6.
The capacity for engineering ketocarotenoid biosynthesis expands the product range which can be achieved from this easy-to-handle organism, with CrBKT expression producing canthaxanthin and the combination of CrBKT + CrCHYB astaxanthin. Our results indicate that despite a subtle reduction in overall growth rates when cells produce ketocarotenoids (Figs. 4 and 5), they are still amenable to cultivation in Fig. 6. Extractable products from wild-type and engineered C. merolae 10D. The schematic displays the extractable products that can be obtained from C. merolae 10D cells through various extraction phases. The dotted arrows indicate the carotenoid fractions that can be extracted from the corresponding cell lines: WT, ii (CrBKT), and viii (CrBKT + CrCHYB). The pigment fractions are named based on the predominant carotenoid present in the extract: ZX (zeaxanthin), CX (canthaxanthin), and ASX (astaxanthin). Additionally, phycocyanin (PC) and chlorophyll a (CHL) are present in all lines. Pigments were extracted as described in M&M and ~1 mL of each was photographed in 3 mL cuvettes. extreme conditions and do not reduce their overall carotenoid contents, even in high light conditions. Future optimization of cultivation parameters can tease-apart the best light and temperature regimes to promote biomass accumulation and increase cellular classes of photopigments in the engineered cells. Our work indicates that C. merolae 10D could be cultivated outdoors, even in some of the hottest urban environments in the world during summer months, but techno-economic analysis would be required to determine whether the CAPEX required to build a controlled bioreactor with constant illumination would be more beneficial than simply using outdoor environmental conditions in situ. This is also encouraged by the recent finding that C. merolae can be adapted to be grown in acidified sea water salinities, further expanding its possible range of geographical application (Hirooka et al., 2020;Villegas-Valencia et al., 2023). Indeed, each implementation of such a cultivation would require individual case-considerations. The thermal extreme tolerance of C. merolae 10D and its engineered derivatives at least suggests that cooling will not be needed if bioreactors are placed outdoors. Waste-heat may be used to optimize culture conditions, especially during colder seasons, as this is energetically less challenging to engineer into a cultivation apparatus than cooling in these extreme environments.

Conclusions and outlook
Here, we show the power of in silico design and de novo construction of transgene expression constructs in an emerging host microalga. We used these molecular tools to rapidly demonstrate the production of nonnative ketocarotenoids in the polyextremohilic red microalga which has emerged in recent years as a promising alternative to other green algal hosts. This work represents the first demonstration of carotenoid metabolic engineering by recombinant technologies in any red alga. The lack of impact alternative carotenoid production had on water-soluble phycocyanin contents adds interesting value to an already specialized algal biomass as these products can be separately extracted as watersoluble and -insoluble fractions from the biomass. The wild-type strain is already a source of zeaxanthin, and our findings indicate it is possible to tailor this host into a production vehicle for either canthaxanthin or astaxanthin without contaminating alpha carotenoids. Given each of these carotenoids has a value of their own, parallel cell lines could be used to generate multiple products from the same algal cultivation infrastructure. Adaptability to saline conditions, high temperature tolerance, and the capacity for growth on high strength waste-waters also encourage the potential value economics of C. merolae bioproduction processes. Given the relative ease of transgene integration into the nuclear genome of this alga and high expression rates, it will likely rapidly become a host cell for a range of photosynthetic engineering concepts.

Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability
All data used in this manuscript can be found in the supplemental files.